The Syntax of Disfluency in Spontaneous Spoken Language
This paper describes a project to parse transcripts of spontaneous spoken dialogues. It concentrates on the issue of disfluency -- that is pauses, fillers (`um's and `er's), repetitions, speech repairs, and fresh starts -- which makes spontaneous speech different from written language.
If one wants to parse spontaneous spoken language, then one needs
some way to cope with disfluency. The approach taken here is that
one should attempt to parse disfluencies along with the fluent
speech. This paper looks at some possible methods of augmenting a
phrase structure grammar to deal with disfluent speech. Such an
augmented grammar was implemented in a parser which was used to
parse a corpus. The results of this parse are reported and
(May 1998: 24 pages)
Ref. No. HCRC/RP-95 Price: UKL ???
Padraic Monaghan and Keith Stenning
Learning to solve syllogisms by watching others' learning
Students learn vicariously when watching other students in a learning
situation. Exactly how and what students learn when watching others,
however, is underexplored. This study reports an experiment on
students learning to solve syllogisms using Euler's Circles by viewing
videos of students learning to use the method. Two conditions were
employed: one group watched a student who showed difficulty in using
the method, the other group watched a student who used the method with
ease. Groups that learned vicariously paced themselves in accordance
with the speed of solution of the students they viewed, the control
group worked much faster at the problems. The different groups seemed
to learn different skills from watching other students learn: watching
a student struggle assisted with manipulating the graphical
representations, watching a student use the method with ease
benefitted translation into and out of the representations.
(October 1998: 13 pages)
Ref. No. HCRC/RP-98 Price: UKL ???